Particle filtering algorithm based on recursive bayesian estimation using gaussian sum in WSN

被引：0

作者：

Dong, Wen ^{[1
]}

Fang, Xiang ^{[1
]}

Chen, Zhiyang ^{[1
]}

Zhang, Weiping ^{[1
]}

机构：

[1] Engineering Institute of Engineer Corps, PLA University of Science and Technology, Jiangsu, Nanjing 210007, China

来源：

Journal of Information and Computational Science | 2012年 / 9卷 / 02期

关键词：

Sensor nodes - State estimation - Monte Carlo methods - Gaussian distribution - Power management;

D O I：

暂无

中图分类号：

学科分类号：

摘要：

To solve the problem of the particle filter algorithm with high computation load and centralized degree, and Gaussian sum recursive Bayesian estimation occurs mixing collapse especially in large process noise environment, a distributed and energy-aware particle filter algorithm is proposed. In the beginning, Gaussian sum recursive Bayesian is used for state estimation to reduce energy. When the estimation error covariance is reduced to δ2, algorithm carries out particle filter and Gaussian sum model is presented to update the particles. Approximated posteriori density distribution replace the re-sampling step by Gaussian sum model in order to enhance the variety of re-sampling particles; The next execution node is selected for the distributed processing by calculating the Kullback-leibler (KL) divergent upper-bound of each neighbor node. Finally, the performance of this algorithm by computer simulation is demonstrated. Copyright © 2012 Binary Information Press.

引用

页码：377 / 386